# Truth Bear (remote · api.truthbear.co)

Verifiable official-source records for AI agents: record_hash + Ed25519 citation, x402 per-call

- Trust score: 78/100 (medium)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- remote · `api.truthbear.co`: 78/100 (this document), [markdown](https://verifymcp.io/servers/changchinfu-truthbear-mcp/api.md), [page](https://verifymcp.io/servers/changchinfu-truthbear-mcp/api)

## Channel facts

- Endpoint: `https://api.truthbear.co/mcp`
- Transports: `streamable-http`
- Auth: `none`
- Version: `1.0.0`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-21.

- **Endpoint Security**: 80/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - No authorisation is required to call this server. Nothing on it is marked destructive, but we can't show that an anonymous caller can only read: full credit needs the server to list at least one tool and every tool to be verifiably read-only. We use a best-effort check of tool names and descriptions to tell read-only tools from tools that change things, so some genuinely read-only tools may not get full credit. Names that start with a read verb (get, list, read, search and similar) and accurate readOnlyHint annotations help. Meet that to earn full credit, or require authorisation.
  - HTTPS is enforced; there's no plaintext access path.
  - The HSTS (Strict-Transport-Security) header is present.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 69/100
  - AI-judged instruction clarity (excellent).
  - Context-footprint check failed: tool/resource definitions use about 1009 tokens (~201/item across 5 items; 5 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 43/100
  - Stability observed for 13 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 5 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 6 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### How do I install the Truth Bear MCP server?

Truth Bear is a hosted endpoint at https://api.truthbear.co/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add --transport http changchinfu-truthbear-mcp 'https://api.truthbear.co/mcp'
```

### Cursor

```json
{
  "mcpServers": {
    "changchinfu-truthbear-mcp": {
      "url": "https://api.truthbear.co/mcp"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "changchinfu-truthbear-mcp": {
      "type": "http",
      "url": "https://api.truthbear.co/mcp"
    }
  }
}
```

### Codex

```toml
[mcp_servers.changchinfu-truthbear-mcp]
url = "https://api.truthbear.co/mcp"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "changchinfu-truthbear-mcp": {
      "type": "remote",
      "url": "https://api.truthbear.co/mcp",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add changchinfu-truthbear-mcp --url 'https://api.truthbear.co/mcp' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  changchinfu-truthbear-mcp:
    url: "https://api.truthbear.co/mcp"
```

### Netclaw

```json
{
  "McpServers": {
    "changchinfu-truthbear-mcp": {
      "Transport": "http",
      "Url": "https://api.truthbear.co/mcp"
    }
  }
}
```

### Vellum

```bash
assistant mcp add changchinfu-truthbear-mcp -t streamable-http -u 'https://api.truthbear.co/mcp'
```

### Other

```json
{
  "mcpServers": {
    "changchinfu-truthbear-mcp": {
      "type": "http",
      "url": "https://api.truthbear.co/mcp"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-19 (score 78, +1)

No change was recorded against any check on this day. Stability & Change Management went from 33 to 37. That category is still filling its 30-day observation window: 10 days of observed history at the previous scan, 11 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-17 (score 77, +1)

No change was recorded against any check on this day. Stability & Change Management went from 27 to 30. That category is still filling its 30-day observation window: 8 days of observed history at the previous scan, 9 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-15 (score 76, +1)

No change was recorded against any check on this day. Stability & Change Management went from 20 to 23. That category is still filling its 30-day observation window: 6 days of observed history at the previous scan, 7 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-13 (score 75, +1)

No change was recorded against any check on this day. Stability & Change Management went from 13 to 17. That category is still filling its 30-day observation window: 4 days of observed history at the previous scan, 5 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-11 (score 74, +1)

No change was recorded against any check on this day. Stability & Change Management went from 7 to 10. That category is still filling its 30-day observation window: 2 days of observed history at the previous scan, 3 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-09 (score 73, +1)

- [functional improvement] Stability: unverified → 0.03

### 2026-11-08 (score 72)

First indexed and scored.

## MCP tools (5)

### `verify_citation` (~107 tokens)

Verify a Truth Bear record (free)

FREE. Given a record_hash from any Truth Bear record, look it up and recompute the canonical hash server-side, returning whether it is a genuine Truth Bear official record plus a plain-language reverse lookup of what exactly that hash attests. Use this to check a citation you were handed by another agent or document BEFORE relying on it. No payment, no API key.

Input parameters:

- `record_hash` (string, required): sha256:<64 hex>, or a >=8-hex prefix as posted publicly

Output parameters:

- `endpoint` (string): Always "gauge/verify".
- `entity` (string|null): Object id the reading is about.
- `error` (string): Present only when the call failed; human-readable reason.
- `found` (boolean): Whether a published Truth Bear reading matches this hash.
- `hash_core_version` (integer|null): Version of the canonical hash core used.
- `how_to_check_a_claimed_value` (string): Present when found: how to check a value someone claims.
- `is_truth_bear_record` (boolean|null): true / false, or null when the check could not be run (see undetermined).
- `note` (string): Screening-level disclaimer.
- `ots_status` (string|null): Per-record field, unused: individual records are not stamped one by one. Anchoring happens per write-batch - see GET /gauge/anchor?hash= for the Merkle inclusion proof and, once the calendar confirms…
- `recomputed_hash` (string|null): sha256:<64 hex> recomputed from the stored record.
- `record_status` (object|null): Three-state disclosure. state="active" for a normal record; a non-active state (e.g. "quarantined") means the record is genuine and its hash still verifies, but we have since flagged it as not fit fo…
- `reverse_lookup` (string): Plain-language statement of exactly what this hash attests.
- `scope` (string): What this hash does and does not attest.
- `signal_id` (string|null): e.g. hydrology.river-level
- `snapshot_date` (string|null): YYYY-MM-DD of the reading.
- `source_ref` (string|null): URL of the official source record.
- `undetermined` (boolean): true when the lookup itself failed (e.g. DB unavailable) - NOT a verdict of "fake".
- `verified` (boolean): Whether the canonical hash recomputed server-side matches the one supplied.

### `find_signal` (~135 tokens)

Find which signals/entities are covered (free)

FREE coverage + freshness manifest: which signal_id lines exist, how many entities each covers, and fresh/recent/stale counts - so you can check "is my entity covered and how fresh" BEFORE paying. Optional filters: industry, signal_id, entity.

Input parameters:

- `entity` (string): object id, e.g. a USGS site id
- `full` (boolean): true = per-entity detail (larger); default compact summary
- `industry` (string): restrict to one domain, e.g. hydrology, agriculture, energy
- `signal_id` (string): e.g. hydrology.river-level

Output parameters:

- `cache_note` (string): What the cache does and does not affect.
- `cache_policy` (string): Cache policy in force.
- `cache_ttl_s` (number): Cache lifetime of this manifest in seconds.
- `detail` (string): Present on some failures: the upstream endpoint's own error text.
- `error` (string): Present only when the call failed; human-readable reason.
- `freshness_rule` (string): How fresh/recent/stale are decided.
- `freshness_thresholds` (object): The numeric thresholds behind freshness_rule.
- `nature` (string): What this manifest is.
- `note` (string): Screening-level disclaimer.
- `signals` (array): One entry per signal line. Compact by default; pass full=true for per-entity detail.
- `totals` (object): signal_ids / industries / distinct_entities / freshness counts across the whole catalogue.
- `update_status_rule` (object): How on_schedule / overdue / not_published / unknown are decided.

### `get_official_record` (~304 tokens)

Get the x402 payment challenge for a paid record

Returns the REAL x402 payment challenge (accepts[]: network / asset / payTo / amount) for the paid endpoint that serves a given signal_id+entity, or any listed endpoint addressed directly by `path`. This tool does NOT deliver paid data and does NOT take payment - MCP has no payment layer. Pay at the returned url with your own x402 client (USDC on Base or Solana, gasless EIP-3009; pay from a plain EOA) and you receive the record directly, with record_hash you can verify offline. This tool only fetches the 402 challenge; it delivers no paid data and collects no payment.

Input parameters:

- `dim` (string): which product: full read (gauge) or a single add-on; the price of each is in the 402 challenge this tool returns, never here
- `entity` (string): which object within that line, e.g. the USGS site id 07010000
- `path` (string): Alternative addressing: any listed endpoint path, optionally with its query string (e.g. "/gauge/chokepoint-region?loc=hormuz"). The full list of legal paths is the free /manifest?all=1. Use this for…
- `signal_id` (string): which signal line, e.g. hydrology.river-level - use find_signal to list them

Output parameters:

- `after_payment` (string): What the paid response carries and how to verify it offline.
- `detail` (string): Present on some failures: the upstream endpoint's own error text.
- `error` (string): Present only when the call failed; human-readable reason.
- `expected` (string): Present on failure: what was expected (an HTTP 402 challenge).
- `got` (integer): Present on failure: the HTTP status actually returned.
- `hint` (string): Present on failure: the next thing to try, e.g. find_signal.
- `how_to_pay` (string): How to sign and send the X-PAYMENT header; pay from a plain EOA.
- `method` (string): HTTP method to use at pay_at (always GET).
- `note` (string): Standing statement that this server never takes payment.
- `pay_at` (string): URL to send the paid GET to.
- `x402_challenge` (object): The upstream 402 body verbatim: x402Version + accepts[] (scheme / network / asset / payTo / amount).

### `purchase_options` (~170 tokens)

How can I actually buy this? (free)

FREE. Ask how to pay for a listed endpoint and get every channel with its REAL status - including the ones that are wired but not open yet. Three channels exist: x402 micropayment (for an agent with a funded USDC wallet on Base), a stateless subscription Bearer token (for a person or team - no wallet, no private key, no account), and a card rail. This tool takes no payment and holds no key; it only tells you which doors exist and which are open today.

Input parameters:

- `path` (string): A listed endpoint path (e.g. "/gauge/grid-reliability-region") to get its price alongside the channels. Omit it to get the channel list on its own. The full list of legal paths is the free /manifest?…

Output parameters:

- `asked_about` (string|null): the path you asked about, or null if you asked for the channel list on its own
- `detail` (string): the upstream error text, when there is one
- `endpoint` (string): always "/buy" - which endpoint answered
- `error` (string): present only when the buy endpoint could not be reached
- `free_first` (string): what stays free and how to avoid paying for a lookup with no data
- `note_unknown_path` (string): present only when asked_about matched no listed endpoint
- `price` (string): the listed price, present only when asked_about matches a listed endpoint
- `title` (string|null): the listing title of that endpoint, when it has one
- `ways_to_buy` (array): every payment channel with its REAL status - including channels that are wired but not open yet

### `verify_triangular` (~220 tokens)

Cross-check a line of official series & recompute the verdict (free)

FREE. Cross-verify one GAUGE interweave line: it returns each official series ("leg") in the line with its own record_hash, current/previous values, expected direction and per-leg minMove deadband, plus a count-style state (directions_matched / direction_mismatch / inverse_consistent / not_comparable / insufficient) and N-of-M. Everything needed to recompute the verdict OFFLINE is in the response - verify each leg hash with verify_citation, then recompute dir=sign(current-previous) with the minMove deadband. Descriptive only: no forecast, no causation, no advice. No payment, no API key.

Input parameters:

- `entity` (string, required): the object within the line, as US-<ST>, e.g. "US-TX" (case-insensitive). Required.
- `line` (string, required): which interweave line, e.g. "E" (state vs national unemployment) or "F" (drought vs crop). Public demo lines are E and F.

Output parameters:

- `detail` (string): Present on some failures: the upstream endpoint's own error text.
- `entity` (string): the external contract code, e.g. US-TX.
- `error` (string): Present only when the call failed; human-readable reason.
- `legs` (array): one entry per official series in the line.
- `line` (string): the line code you asked for, e.g. E or F.
- `matched` (integer): N: how many REQUIRED legs had measured dir === expect this window.
- `nature` (string|null): neutral English disclosure; may be null.
- `note` (string): standing disclaimer: direction is observed, not a trend/forecast/causation claim.
- `pair_state` (string|null): inverse-pair lines (F/FW) only: holds_stress | holds_relief | broken | insufficient | not_comparable; absent/null on other lines.
- `reason` (string|null): stale | missing_current_or_previous | below_min_move | period_misaligned | no_active_drought | parallel_display | null.
- `required` (integer): M: number of required legs. The public claim is exactly "N of M moved in the registered direction".
- `snapshot_window` (object|null): the {from,to} window the legs were read over, or null.
- `state` (string): directions_matched | direction_mismatch | inverse_consistent | not_comparable | insufficient. A count/observation, NOT an adjudication.
- `statement` (string): the locked, neutral disclosure sentence with N/M substituted.
- `title` (string): letter-only title (e.g. "線E"); never a phenomenon name.
- `verify_hint` (string): exactly how to recompute the state offline from each leg (record_hash + dir=sign(Δ) with minMove deadband).

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/changchinfu-truthbear-mcp/api#diagnostics

## Score history

- 2026-11-21: 78
- 2026-11-20: 78
- 2026-11-19: 78
- 2026-11-18: 77
- 2026-11-17: 77
- 2026-11-16: 76
- 2026-11-15: 76
- 2026-11-14: 75
- 2026-11-13: 75
- 2026-11-12: 74
- 2026-11-11: 74
- 2026-11-10: 73
- 2026-11-09: 73
- 2026-11-08: 72

## Common questions

### What is the Truth Bear MCP server?

Truth Bear is an MCP server listed in the public MCP registry as io.github.CHANGCHINFU/truthbear-mcp. Verifiable official-source records for AI agents: record_hash + Ed25519 citation, x402 per-call. This page covers its hosted endpoint (https://api.truthbear.co/mcp).

### Is the Truth Bear MCP server safe to use?

Truth Bear scores 78 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the Truth Bear MCP server expose?

Truth Bear exposes 5 tools: verify_citation, find_signal, get_official_record, purchase_options, verify_triangular. Their descriptions and schemas cost roughly 936 tokens of context every time the server is loaded.

### Does the Truth Bear MCP server require authentication?

No. We connected to Truth Bear without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

### Is the Truth Bear MCP server still maintained?

Truth Bear is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What did VerifyMCP test on the Truth Bear MCP server?

VerifyMCP ran 13 checks on Truth Bear, covering TLS certificate, Authorization, HTTPS, HSTS header and 9 more. Supply Chain Security and Provenance & Transparency do not apply to a hosted endpoint, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- Remote endpoint: https://api.truthbear.co/mcp
- Repository: https://github.com/CHANGCHINFU/truthbear-mcp
- Changelog RSS feed: https://verifymcp.io/servers/changchinfu-truthbear-mcp/api.xml
- Changelog JSON feed: https://verifymcp.io/servers/changchinfu-truthbear-mcp/api/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/changchinfu-truthbear-mcp/api.json
- HTML version of this page: https://verifymcp.io/servers/changchinfu-truthbear-mcp/api
